Identification of extracellular vesicles from their Raman spectra via self-supervised learning.

Extracellular vesicles (EVs) released from cells attract interest for their possible role in health and diseases. The detection and characterization of EVs is challenging due to the lack of specialized methodologies. Raman spectroscopy, however, has been suggested as a novel approach for biochemical analysis of EVs. To extract information from the spectra, a novel deep learning architecture is explored as a versatile variant of autoencoders. The proposed architecture considers the frequency range separately from the intensity of the spectra. This enables the model to adapt to the frequency range, rather than requiring that all spectra be pre-processed to the same frequency range as it was trained on. It is demonstrated that the proposed architecture accepts Raman spectra of EVs and lipoproteins from 13 biological sources and from two laboratories. High reconstruction accuracy is maintained despite large variances in frequency range and noise level. It is also shown that the architecture is able to cluster the biological nanoparticles by their Raman spectra and differentiate them by their origin without pre-processing of the spectra or supervision during learning. The model performs label-free differentiation, including separating EVs from activated vs. non-activated blood platelets and EVs/lipoproteins from prostate cancer patients versus non-cancer controls. The differentiation is evaluated by creating a neural network classifier that observes the features extracted by the model to classify the spectra according to their sample origin. The classification reveals a test sensitivity of 92.2 % and selectivity of 92.3 % over 769 measurements from two labs that have different measurement configurations.

Extracellular vesicles from activated platelets possess a phospholipid-rich biomolecular profile and enhance prothrombinase activity.

BACKGROUND: Extracellular vesicles (EVs), in particular those derived from activated platelets, are associated with a risk of future venous thromboembolism. OBJECTIVES: To study the biomolecular profile and function characteristics of EVs from control (unstimulated) and activated platelets. METHODS: Biomolecular profiling of single or very few (1-4) platelet-EVs (control/stimulated) was performed by Raman tweezers microspectroscopy. The effects of such EVs on the coagulation system were comprehensively studied. RESULTS: Raman tweezers microspectroscopy of platelet-EVs followed by biomolecular component analysis revealed for the first time 3 subsets of EVs: (i) protein rich, (ii) protein/lipid rich, and (iii) lipid rich. EVs from control platelets presented a heterogeneous biomolecular profile, with protein-rich EVs being the main subset (58.7% ± 3.5%). Notably, the protein-rich subset may contain a minor contribution from other extracellular particles, including protein aggregates. In contrast, EVs from activated platelets were more homogeneous, dominated by the protein/lipid-rich subset (>85%), and enriched in phospholipids. Functionally, EVs from activated platelets increased thrombin generation by 52.4% and shortened plasma coagulation time by 34.6% ± 10.0% compared with 18.6% ± 13.9% mediated by EVs from control platelets (P = .015). The increased procoagulant activity was predominantly mediated by phosphatidylserine. Detailed investigation showed that EVs from activated platelets increased the activity of the prothrombinase complex (factor Va:FXa:FII) by more than 6-fold. CONCLUSION: Our study reports a novel quantitative biomolecular characterization of platelet-EVs possessing a homogenous and phospholipid-enriched profile in response to platelet activation. Such characteristics are accompanied with an increased phosphatidylserine-dependent procoagulant activity. Further investigation of a possible role of platelet-EVs in the pathogenesis of venous thromboembolism is warranted.

Cancer cell line-specific protein profiles in extracellular vesicles identified by proteomics.

Extracellular vesicles (EVs), are important for intercellular communication in both physiological and pathological processes. To explore the potential of cancer derived EVs as disease biomarkers for diagnosis, monitoring, and treatment decision, it is necessary to thoroughly characterize their biomolecular content. The aim of the study was to characterize and compare the protein content of EVs derived from three different cancer cell lines in search of a specific molecular signature, with emphasis on proteins related to the carcinogenic process. Oral squamous cell carcinoma (OSCC), pancreatic ductal adenocarcinoma (PDAC) and melanoma brain metastasis cell lines were cultured in CELLine AD1000 flasks. EVs were isolated by ultrafiltration and size-exclusion chromatography and characterized. Next, the isolated EVs underwent liquid chromatography-mass spectrometry (LC-MS) analysis for protein identification. Functional enrichment analysis was performed for a more general overview of the biological processes involved. More than 600 different proteins were identified in EVs from each particular cell line. Here, 14%, 10%, and 24% of the identified proteins were unique in OSCC, PDAC, and melanoma vesicles, respectively. A specific protein profile was discovered for each cell line, e.g., EGFR in OSCC, Muc5AC in PDAC, and FN1 in melanoma vesicles. Nevertheless, 25% of all the identified proteins were common to all cell lines. Functional enrichment analysis linked the proteins in each data set to biological processes such as "biological adhesion", "cell motility", and "cellular component biogenesis". EV proteomics discovered cancer-specific protein profiles, with proteins involved in processes promoting tumor progression. In addition, the biological processes associated to the melanoma-derived EVs were distinct from the ones linked to the EVs isolated from OSCC and PDAC. The malignancy specific biomolecular cues in EVs may have potential applications as diagnostic biomarkers and in therapy.

Proteomic and histopathological characterisation of sicca subjects and primary Sjögren's syndrome patients reveals promising tear, saliva and extracellular vesicle disease biomarkers.

BACKGROUND: Mononuclear cell infiltration of exocrine glands, production of Ro/SSA and La/SSB autoantibodies, along with oral and ocular dryness, are characteristic features of primary Sjögren's syndrome (pSS). Non-SS sicca subjects, an underexplored group in relation to pSS, display similar sicca symptoms, with possible mild signs of inflammation in their salivary glands, yet with no serological detection of autoantibody production. In this study, we investigated inflammatory manifestations in the salivary gland tissue, tear fluid and saliva of non-SS subjects, as compared to pSS patients and healthy individuals. METHODS: Fifteen non-SS, 10 pSS and 10 healthy subjects were included in the analyses. Histological evaluation of salivary gland biopsies was performed. Liquid chromatography-mass spectrometry (LC-MS) was conducted on tear fluid and stimulated whole saliva, and proteomic biomarker profiles were generated. Extracellular vesicle (EVs) isolation and characterisation from both fluids were also combined with LC-MS. The LC-MS data were analysed for quantitative differences between patient and control groups using Scaffold. Database for Annotation, Visualization and Integrated Discovery (DAVID) and Functional Enrichment Analysis Tool (FunRich) were applied for functional analyses. RESULTS: Histopathological evaluation of salivary gland biopsies showed implications of milder inflammation in non-SS subjects through mononuclear cell infiltration, fibrosis and fatty replacement, as compared to pSS patients. Although unaffected in the non-SS group, upregulation of proinflammatory pathways and proteins involved in ubiquitination (LMO7 and HUWE1) and B cell differentiation (TPD52) were detected in tear fluid of pSS patients. Moreover, overexpression of proteins STOM, ANXA4 and ANXA1, regulating cellular innate and adaptive immunological pathways, were further identified in EVs from tear fluid of pSS patients. Finally, whole saliva and EVs isolated from whole saliva of pSS patients expressed proteins vital for innate MHC class I cellular regulation (NGAL) and T cell activation (CD44). CONCLUSIONS: Non-SS sicca subjects may show implications of mild inflammation in their glandular tissue, while their protein profile was strikingly more similar to healthy controls than to pSS patients. Hence, the tear and salivary biomarkers identified could be implemented as potential non-invasive diagnostic tools that may aid in increasing diagnostic accuracy when evaluating non-SS subjects and pSS patients and monitoring disease progression.

Efficient extracellular vesicle isolation by combining cell media modifications, ultrafiltration, and size-exclusion chromatography.

Extracellular vesicles (EVs) are a heterogeneous population of biological particles released by cells. They represent an attractive source of potential biomarkers for early detection of diseases such as cancer. However, it is critical that sufficient amounts of EVs can be isolated and purified in a robust and reproducible manner. Several isolation methods that seem to produce distinct populations of vesicles exist, making data comparability difficult. While some methods induce cellular stress that may affect both the quantity and function of the EVs produced, others involve expensive reagents or equipment unavailable for many laboratories. Thus, there is a need for a standardized, feasible and cost-effective method for isolation of EVs from cell culture supernatants. Here we present the most common obstacles in the production and isolation of small EVs, and we suggest a combination of relatively simple strategies to avoid these. Three distinct cell lines were used (human oral squamous cell carcinoma (PE/CA-PJ49/E10)), pancreatic adenocarcinoma (BxPC3), and a human melanoma brain metastasis (H3). The addition of 1% exosome-depleted FBS to Advanced culture media enabled for reduced presence of contaminating bovine EVs while still ensuring an acceptable cell proliferation and low cellular stress. Cells were gradually adapted to these new media. Furthermore, using the Integra CELLine AD1000 culture flask we increased the number of cells and thereby EVs in 3D-culture. A combination of ultrafiltration with different molecular weight cut-offs and size-exclusion chromatography was further used for the isolation of a heterogeneous population of small EVs with low protein contamination. The EVs were characterized by nanoparticle tracking analysis, immunoaffinity capture, flow cytometry, Western blot and transmission electron microscopy. We successfully isolated a significant amount of small EVs compatible with exosomes from three distinct cell lines in order to demonstrate reproducibility with cell lines of different origin. The EVs were characterized as CD9 positive with a size between 60-140 nm. We conclude that this new combination of methods is a robust and improved strategy for the isolation of EVs, and in particular small EVs compatible with exosomes, from cell culture media without the use of specialized equipment such as an ultracentrifuge.

Identification of potential saliva and tear biomarkers in primary Sjögren's syndrome, utilising the extraction of extracellular vesicles and proteomics analysis.

BACKGROUND: There is a long-lasting need for non-invasive, more accurate diagnostic techniques when evaluating primary Sjögren's syndrome (pSS) patients. Incorporation of additional diagnostics involving screening for disease-specific biomarkers in biological fluid is a promising concept that requires further investigation. In the current study we aimed to explore novel disease biomarkers in saliva and tears from pSS patients. METHODS: Liquid chromatography-mass spectrometry (LC-MS) was performed on stimulated whole saliva and tears from 27 pSS patients and 32 healthy controls, and salivary and tear proteomic biomarker profiles were generated. LC-MS was also combined with size exclusion chromatography to isolate extracellular vesicles (EVs) from both fluids. Nanoparticle tracking analysis was conducted on joint fractions from the saliva and tears to determine size distribution and concentration of EVs. Further EV characterisation was performed by immunoaffinity capture of CD9-positive EVs using magnetic beads, detected by flow cytometry. The LC-MS data were analysed for quantitative differences between patient and control groups using Scaffold, and the proteins were further analysed using the Database for Annotation, Visualization and Integrated Discovery (DAVID), for gene ontology overrepresentation, and the Search Tool for the Retrieval of Interacting Genes/Proteins for protein-protein interaction network analysis. RESULTS: Upregulation of proteins involved in innate immunity (LCN2), cell signalling (CALM) and wound repair (GRN and CALML5) were detected in saliva in pSS. Saliva EVs also displayed biomarkers critical for activation of the innate immune system (SIRPA and LSP1) and adipocyte differentiation (APMAP). Tear analysis indicated overexpression of proteins involved in TNF-α signalling (CPNE1) and B cell survival (PRDX3). Moreover, neutrophil gelatinase-associated lipocalin was upregulated in saliva and tears in pSS. Consistently, DAVID analysis demonstrated pathways of the adaptive immune response in saliva, of cellular component assembly for saliva EVs, and of metabolism and protein folding in tears in pSS patients. CONCLUSIONS: LC-MS of saliva and tears from pSS patients, solely and in combination with size-exclusion chromatography allowed screening for possible novel biomarkers encompassing both salivary and lacrimal disease target organs. This approach could provide additional diagnostic accuracy in pSS, and could possibly also be applied for staging and monitoring the disease.